style-standards

bsamiee's avatarfrom bsamiee

Enforces style consistency for documentation and code with support for taxonomy, voice, and formatting domains. Use when creating or modifying files requiring: (1) markdown structure (headers, lists, tables, Dictums, separators), (2) code organization (comments, section dividers, naming conventions), (3) voice compliance (tone, grammar, imperative phrasing), or (4) formatting validation against project standards.

⭐1stars🔀0forks📁View on GitHub🕐Updated Dec 27, 2025

When & Why to Use This Skill

This Claude skill automates the enforcement of professional style standards across documentation and codebases. It ensures structural integrity, voice consistency, and formatting precision, specifically optimized for LLM comprehension and agentic workflows by reducing token noise and maximizing instruction clarity.

Use Cases

  • Standardizing technical documentation (Markdown) to ensure consistent header hierarchies, table structures, and professional terminology.
  • Enforcing code organization rules, including naming conventions that prohibit vague terms like 'utils' or 'manager' and front-loading architectural intent in comments.
  • Optimizing agentic instructions and prompts by applying 'Voice' standards such as active voice and imperative phrasing to improve LLM execution accuracy.
  • Validating project-wide formatting in monorepos, ensuring that whitespace, separators, and document depth adhere to strict technical standards for better maintainability.
  • Streamlining content creation for developers by providing a 'Quality Gate' that checks for grammar, tone, and structural compliance before file finalization.
namestyle-standards
typestandard
depthextended
descriptionEnforces style consistency for documentation and code with support for taxonomy, voice, and formatting domains. Use when creating or modifying files requiring: (1) markdown structure (headers, lists, tables, Dictums, separators), (2) code organization (comments, section dividers, naming conventions), (3) voice compliance (tone, grammar, imperative phrasing), or (4) formatting validation against project standards.

[H1][STYLE-STANDARDS]

Dictum: Style consistency maximizes agent comprehension.


Govern file creation and modification in monorepo.

Tasks:

  1. Read index.md — Reference file listing for navigation
  2. Read keywords.md — Canonical keyword list; all Markers use official terms
  3. (taxonomy) Read taxonomy.md — Lexicon, references, stati definitions
  4. (voice) Read voice.md — Grammar, ordering, comments, constraints, naming, density
  5. (formatting) Read formatting.md — Structure, typeset, spacing, examples
  6. Apply standards — Implement per domain Guidance and Best-Practices
  7. Validate — Quality gate; see §VALIDATION

Scope:

  • Documentation: Markdown structure, headers, lists, tables, Dictums, separators.
  • Code: Comments, headers, section dividers, naming conventions, file organization.

Domain Navigation:

  • [TAXONOMY] — Terms, markers, cross-references. Load for: sigils, stati, lexicon definitions.
  • [VOICE] — Tone, grammar, comments, naming. Load for: imperative phrasing, headers, code naming.
  • [FORMATTING] — Layout, separators, spacing. Load for: header structure, dividers, whitespace rules.

[REFERENCE]: index.md — Reference file listing


[1][TAXONOMY]

Dictum: Vocabulary anchors structure; Markers encode state.


Signals intent for agent execution. Leverage terms for document traversal.

Guidance:

  • Dictum - Read Dictum + headers first—rapid file mapping.
  • Qualifier - [ALWAYS] respect inline directives when encountered.
  • Preamble - Signals section-wide imperative.
  • Terminus - Signals task-specific imperative; isolated effect.
  • Corpus - Read after Preamble/Terminus orientation.
  • Gate - [CRITICAL] Finalize checklist items prior to proceeding; use [VERIFY] for Gate checklists.
  • Directive - Lists require strict adherence; polarity set by Modifier.
  • Stati - Replace emoji.

Best-Practices:

  • Markers: Hard limit: 10 per file. Strategic placement maximizes compliance.
    • Preamble/Terminus - 0–4 markers per file maximum.

[REFERENCE]: →taxonomy.md — Lexicon, references, stati


[2][VOICE]

Dictum: Universal standards for LLM-optimized context, documentation, and agentic instructions.


Applies to documentation and comments. Scope: tone, list semantics, ordering primacy, grammar, syntax, modals, visuals, comment standards, keywords.

Guidance:

  • Voice - Active voice: 56% token reduction.
  • Tone - Mechanical, domain-specific. No hedging, no self-reference.
  • Syntax - Simple sentences: 93.7% accuracy vs 46.8% nested.
  • Punctuation - Attention sinks—absorb 20-40% weight despite minimal semantic content.
  • Ordering - [CRITICAL] Primacy effects peak at 150-200 instructions; 5.79× attention for early items.
    • Critical-First - Highest-priority constraints at sequence start.
    • Middle Burial - Middle positions suffer U-shaped attention loss.
  • Comments - Front-load architectural decisions where attention peaks.
  • Density - Tables: >2 entities, >2 dimensions. Diagrams: >3 steps or >2 hierarchy levels.

Best-Practices:

  • Comments - Incorrect: 78% accuracy loss—omit if uncertain. Why > What: intent = signal, logic = noise.
  • Constraints - 6+ simultaneous: <25% satisfaction. Max 3-5 per level.
  • Delimiters - Consistency over choice. 18-29% variance per change.
  • Stopwords - Remove the, a, an, please, kindly.
  • Tone - Actions: imperative. Context/facts: declarative.
  • Naming - Prohibited: utils, helpers, misc, config, cfg, opts, params, Data, Info, Manager, Service.

[REFERENCE]: →voice.md — Grammar, ordering, comments, constraints, naming, density


[3][FORMATTING]

Dictum: Whitespace and separator rules for document structure.


Separators encode hierarchy. Whitespace: semantic, not cosmetic. Patterns enable rapid reference.

Guidance:

  • Dictum - Place first after H1/H2. State WHY, not WHAT. Format: >**Dictum:** *statement*
  • Depth - H1: File Truth. H2: Smallest agent read unit. H3: Nesting limit. [CRITICAL] H4+ requires new file.
  • Lists - Use numbered 1. for sequence/priority. Use bullet - for equivalence/sets.
  • Labels - Format parent: **Bold:** with colon. Format child: *Italic:* for contrast.
  • Separators - Use --- for hard boundaries (H2 → H2, H3 → H3). Use <br> for soft transitions (H2 → H3).
  • Spacing - Place 1 blank after header. Place none after ---. Place none between list items.
  • Dividers - Pad code separators // --- [LABEL] --- to column 80.
  • Tables - Include [INDEX] first column. Format headers as [HEADER] sigil. Align: center index, right numeric, left prose.

Best-Practices:

  • Separator Prohibitions - --- between H2 and first H3 prohibited. <br> between sibling H3s prohibited.
  • List Prohibitions - Single-item lists prohibited—use prose. Bullet - only; */+ prohibited. Parallel grammar required.
  • Header Integrity - Level skipping prohibited. H1 → H2 → H3 strictly sequential.
  • Thresholds - Lists: 2-7 items. Items: <100 chars. Nesting: 2 levels max.
  • Sigils - UPPERCASE, max 3 words, underscores for compound. Exception: .claude/ infrastructure (skills, commands, agents) use hyphens matching file/folder name.
  • Soft Breaks - <br> required after Dictum and Preamble. Groups 2-3 related definitions inline.
  • Case Taxonomy - UPPERCASE: sigils, rubrics, keywords, section labels. Title Case: table cells. kebab-case: files.
  • Directive Ordering - [IMPORTANT]: precedes [CRITICAL]:. Within list: [ALWAYS] precedes [NEVER].
  • Table Styling - First column bold for category anchoring.

[REFERENCE]: →formatting.md — Structure, typeset, spacing, example


[4][VALIDATION]

Dictum: Gates prevent non-compliant output.


[VERIFY] Completion:

  • Structure: Nesting ≤H3, critical constraints at sequence start.
  • Voice: Active voice, no stopwords, no hedging, no self-reference.
  • Formatting: Separators correct, spacing rules applied.
  • Consistency: All markers use canonical keywords.

[REFERENCE] Operational checklist: →validation.md